A method and device for detecting total nitrogen in soil

Abstract

The invention provides a soil total nitrogen detection method and a soil total nitrogen detection method device. The soil total nitrogen detection method comprises the following step: based on the spectral reflectivity of soil to be detected in different wavebands under irradiation of a near-infrared light source, through a soil total nitrogen predicting model established by using a gray correlation analyzing and limit learning machine, acquiring the total nitrogen content of the soil to be detected. Through the soil total nitrogen detection method and the soil total nitrogen detection methoddevice, provided by the invention, the gray correlation analyzing and limit learning machine is used for effectively selecting a sensitive waveband and establishing correlation analysis, and the soiltotal nitrogen predicting model with high reliability is established, so that the accuracy of the total nitrogen detection of the soil to be detected is improved; meanwhile, through remote data acquisition, interference of an external environment on a prediction result is effectively avoided and the accuracy of model prediction is further improved.

Description

technical field [0001] The invention relates to the technical field of spectral detection, and more specifically, to a soil total nitrogen detection method and device. Background technique [0002] Soil is the foundation of plant growth and development. The abundance of nutrient elements (nitrogen, phosphorus, potassium, etc.) in the soil directly affects the growth of crops and the formation of yield and quality. Among them, the role of soil nitrogen is mainly to promote the growth of crop roots, stems, The individual growth of leaves is the primary nutrient element to ensure the quality of agricultural products. In order to pursue high yield, excessive fertilization of the soil has caused soil compaction and pollution problems. Therefore, the nitrogen element detection of the soil is carried out first, and then variable fertilization can effectively prevent the excessive use of chemical fertilizers and the occurrence of soil pollution problems. [0003] The spectral analy...

AI Technical Summary

Problems solved by technology

[0004] The above-mentioned method of using support vector machine (SVM) or BP neural network to establish a soil total nitrogen prediction model requires artificially setting a large number of network parameters, which is prone to the problem of local optimal solutions, resulting in poor stability of the model and reducing the quality of the model. For the prediction accuracy of soil total nitrogen content

Images